A device with one or more microphones can enable a speaker (such as a person) to interact with and control part or all of an electronic system, such as one or more appliances on a home automation network. Spoken commands can be detected and processed by an automatic speech recognition system. The accuracy with which a automatic speech recognition system can detect commands can depend on the signal-to-noise ratio of the signal received through the microphone system. A higher signal-to-noise can result in more accurate command detection.